Emanuele Aina authored 4 years ago and Martyn Welch committed 4 years ago

After commit 45dae71b fixed the boot on Renesas boards with
linux >= 5.7 due to the change of the dtb name, the tests exercising
the NFS boot of the current release as the first stage broke since
they do not use `boot_commands` and thus didn't get the override.

To fix that, rather than constructing the path to the DTB from the
LAVA device name, which no longer match the board type with
linux >= 5.7, explicitly pass the DTB path for the first stage and the
release under test separately.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Use the passed URL to Tiny RFS rather than re-construct the URL
with script hardcoded function. This allows to test the RFS prepared
in users GIT branches.

Signed-off-by: Denis Pynkin <denis.pynkin@collabora.com>

Signed-off-by: Benani Sagar Kishore (RBEI/ECO2) <SagarKishore.Benani@in.bosch.com>

Signed-off-by: Ritesh Raj Sarraf <rrs@debian.org>

Signed-off-by: Ritesh Raj Sarraf <rrs@debian.org>

Lightweight CI runners have been introduced to run jobs which have low
resource consumption and do not need privileged containers to reduce costs
and latency, so let's use them.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Add some additional YAML anchors so we further shrink the
*-internal-mx6qsabrelite profile definitions.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Martyn Welch authored 4 years ago and Emanuele Aina committed 4 years ago

Two of the test cases point to videos to show what a successful run looks
like. These were randomly stored on the old wiki machine. Rather than
having unlinked static content on the website, add these to the
`images` folder and update links.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

Commit 822fb8b8 dropped a few testcases, but the profiles still
point to them, resulting in errors like on
https://lava.collabora.co.uk/scheduler/job/2572900

:

    Job error: Unable to open test definition 'test-cases/tracker-indexing-local-storage.yaml':
    [Errno 2] No such file or directory:
    '/var/lib/lava/dispatcher/tmp/2572900/lava-overlay-knf8ynqa/var/lib/lava-2572900/2/tests/21_tracker-indexing-local-storage/test-cases/tracker-indexing-local-storage.yaml'

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

After the discussions following test cases will be dropped.
apparmor-tracker
folks-ofono-backend
telepathy-ring
tracker-indexing-local-storage
tracker-indexing-mass-storage
traffic-control-basic

Apertis: T5734
Signed-off-by: Tarun Baghmar <tarun.baghmar@in.bosch.com>

Denis Pynkin authored 4 years ago and Emanuele Aina committed 4 years ago

Ask to wait a few seconds before to check if undeploy is really
finished. Sometimes the human is too fast to enter commands
so undeploy have no chance to finish before the check.
That is leading to false result of manual test while it works well.

Signed-off-by: Denis Pynkin <denis.pynkin@collabora.com>

Frederic Danis authored 4 years ago and Denis Pynkin committed 4 years ago

wget needs the complete URL to be able to download key0.key file.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

Earlier, we were using an old kerne and rootfs from 18.09.xx.

We've now switched to using a much newer one from v2020, which is where
it requires an update of our assumpt login prompts

Signed-off-by: Ritesh Raj Sarraf <ritesh.sarraf@collabora.com>

In commit 83772e1 and later commits DTB names in Linux kernel was
renamed to align names with config options.
Hence the old DTB name `r8a7796-m3ulcb.dtb` is `8a77960-ulcb.dtb` for
Renesas M3 boards used in LAVA.

Old U-Boot versions are using the old naming schema including version
used in v2021dev3 atm.

This commit set proper name only for v2021dev3 branch allowing to boot
images for Renesas boards in LAVA.

Signed-off-by: Denis Pynkin <denis.pynkin@collabora.com>

Martyn Welch authored 4 years ago and Emanuele Aina committed 4 years ago

Rather than looping over branches (in included template from `pages`),
generate the test cases here, provide these as an artifact and deploy
in a separate job. This avoids the regeneration of pages that have
not changed.

Use the apertis-test-cases-web as part of a multi-project pipeline to
deploy. Remove the deployment tools from this branch.

Whilst we are here, utilise an environment variable for the release to
make it easier to migrate this between releases.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

The bundle generation with an unknown key doesn't display correctly.
Move bundle generation to `pre-conditions`.
Add `cryptsetup` package installation

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

The GitLab CI logic has been moved to the pages branch, remove the
logic from this branch and use the version on the pages branch.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

We want to have the pages updated whenever there's a change on any
`apertis/` branch. Drop the test case and allow the pages pipeline to be
run on any `apertis/` branch.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

We are going to start generating test cases for multiple generations of
Apertis to enable testing of stable releases. Add a menu to enable
browsing of the various releases.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

Having the argument parsing in the atc_renderer code makes it much
harder to use it as a library. Pull the arg parsing into the
application.

As it turns out, atc didn't end up being used in this manner, but fix
it now to avoid this potentially needing to be addressed again in the
future.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

We want to redirect linking that pages without a version specified to
the latest to ensure existing linking continues to work. For everything
else, lets redirect to the main page.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

We want to move to providing test cases for multiple generations of
Apertis so as to enable us to develop the bits we need to do this,
begin by putting the existing output in `latest/`.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

The bundle generation with an unknown key doesn't display correctly.
Add a command to retrieve the `key0.key` file from gitlab.apertis.org.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

By default SabreLite board plays sound through HDMI connector.
In case tester want to output sound through headphone connector he should
configure PulseAudio for it.

After the 2 first commands (set-default-sink and set-sink-port) the sound
is heard correctly, but not after a reboot. This is due to mute state not
saved at this time, this needs the third command (after reboot) to fix it.
Further reboot will play sound correctly.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

This commit add commands to generate a static update bundle with a key
not known by the DUT, from the original encrypted update.
It allows to test that a static update encrypted with an unknown key is not
applied.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

New version of apertis-update-manager expects offline updates to be
encrypted.
The apertis-update-manager-automount test ensure that non encrypted
updates are not applied, while encrypted updates applies correctly.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

Signed-off-by: Ritesh Raj Sarraf <rrs@debian.org>

The ssh-copy-id command syntax in sdk-ade-remote-debugging.yaml is
incorrect in two ways:

- The user is in capitals, whilst the actual username used should be in
  lower case
- The test case uses `$HOST` as a convention for the hostname/IP to be used
  for the target device, this command is missing the dollar sign.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

Baghmar Tarun authored 4 years ago and Martyn Welch committed 4 years ago

Rework the testcase to not require packages to be installed and thus work on systems with an immutable rootfs like the ostree images.

Apertis: T5734

Signed-off-by: Tarun Baghmar <tarun.baghmar@in.bosch.com>

Commit  d0487486 "lava: Reduce redundancy in the profile
definitions" introduced a change that ended up causing issues with the
AUM tests, so let's go back to the previous state.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Emanuele Aina authored 4 years ago and Peter Senna Tschudin committed 4 years ago

Move from the old variable names which dates back to when the pipeline
was on Jenkins to something a bit more consistent.

This would be a breaking change for pipelines which include the
template, but since the main recipe repo has not switched over yet we
can play fast and loose safely.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Emanuele Aina authored 4 years ago and Peter Senna Tschudin committed 4 years ago

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Use YAML anchors and extensions to avoid repeating the same contents over
and over.

This commit does not introduce any change in the generated job
definitions, except for the case below, which is harmless:

    apertis_ostree-minimal-armhf-uboot
    -      commands: ['reset']
    +      commands: ['run bootcmd_mmc0']

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

The test currently fails due to timeout.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

Use the multiple ostree preconditions support introduced in 6b2fd91e
instead of hardcoding the blurb for the second repository.

Also fix the branches and the resulting paths for the unpacked scripts.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

It turns out that LAVA really wants strings for the metadata entries and
didn't like nesting a dictionary there.

For this reason it was failing to submit jobs with this error:

    Submitting job group-apparmor-tpl.yaml: <Fault 400: "Problem with submitted job data: expected str for dictionary value @ data['metadata']['source']">

Unfortunately our custom `lava-submit.py` was silently eating the error
so the pipeline was marked as successful.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Baghmar Tarun authored 4 years ago and Emanuele Aina committed 4 years ago

Rework the testcase to not require packages to be installed and thus
work on systems with an immutable rootfs like the ostree images.

Apertis: T5734

Signed-off-by: Baghmar Tarun <tarun.baghmar@in.bosch.com>

The `${source_wip:=false}` snippet not only sets the `$source_wip`
variable if unset, but also executes the result of its expansion.

In our case it executed `false`, which made the pipeline very unhappy.

However, all tests submitted from this repository directly should be
considered "wip" and it will be up to the image repository including the
gitlab-ci job template to define the logic for which submissions should
be "wip" or not.

So drop the logic and use a plain `source_wip` variable defaulting to
`true`. The meaning of the variable was not documented either, so add a
blurb about it.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>